Supporting information to

Next-generation cophylogeny: unravelling eco-evolutionary processes

Isabel Blasco-Costa1,2, Alexander Hayward3, Robert Poulin4 and Juan A. Balbuena5
1Department of Invertebrates, Natural History Museum of Geneva, PO Box 6434, CH-1211 Geneva 6, Switzerland. ORCID: 0000-0002-5001-5986
2Department of Arctic and Marine Biology, UiT The Arctic University of Norway, Langnes, PO Box 6050, 9037 Tromsø, Norway.
3Centre for Ecology and Conservation, University of Exeter, Penryn Campus, Penryn, Cornwall, Exeter, TR10 9FE, UK. ORCID: 0000-0001-7413-718X
4Department of Zoology, University of Otago, PO Box 56, Dunedin New Zealand. ORCID: 0000-0003-1390-1206
5Cavanilles Institute of Biodiversity and Evolutionary Biology, University of Valencia, PO Box 22085, 46071 Valencia, Spain. ORCID: 0000-0003-4006-1353

We illustrate the application of the Cophylospace approach in R (R Core Team 2021) with an example of 14 Ligophorus spp. (Monogenea: Dactylogyridae) on six Mediterranean and Black Sea species of grey mullets (Mugilidae). The analyses go along the lines of Russo et al. (2017). However, the Mantel tests proposed by these authors to test whether phylogenetically related taxa are more likely to interact with similar partners could not be implemented here. In specialized associations such as often occurs with host and parasites, both closely and distantly related taxa can be equally disimilar in the associations with their partners. So the Mantel tests lack sufficient resolution. In this particular example, Cophylogenetic signal and the interactions between phylogeny and morphology of hosts and parasites will be assessed with the sum of squared residuals (m2) produced by PACo (Balbuena et al. 2013), which is inversely proportional to the concordance between the configurations being tested. Since m2 depends on branch length units, the dissimilarity matrices will be normalized prior to analysis.

The data needed to run this example are embedded in an R object.

LigoMulData <- readRDS("ligophorus_example.rds")

The analyses require three R libraries, phytools (Revell 2012), paco (Hutchinson et al. 2017) and plotly (Sievert 2020).

library(phytools)
## Loading required package: ape
## Warning: package 'ape' was built under R version 4.0.5
## Loading required package: maps
library(paco)
library(plotly)
## Warning: package 'plotly' was built under R version 4.0.5
## Loading required package: ggplot2
## Warning: package 'ggplot2' was built under R version 4.0.4
## 
## Attaching package: 'plotly'
## The following object is masked from 'package:ggplot2':
## 
##     last_plot
## The following object is masked from 'package:stats':
## 
##     filter
## The following object is masked from 'package:graphics':
## 
##     layout

Host Parasite Associations

The tanglegram below depicts the host-parasite associations based on phylogenetic informacion of Ligophorus (González-Rodríguez et al. 2016) and grey mullets (Durand et al. 2012). Data on host-parasite associations is from González-Rodríguez et al. (2017).

Mul.tree <- LigoMulData$mullets_tree
Lig.tree <- LigoMulData$lig_tree
HS <- LigoMulData$assoc_mat
  # Arrange host-parasite associations in pairs 
HS.lut <- which(HS ==1, arr.ind=TRUE)
linkhs <- cbind(rownames(HS)[HS.lut[,1]], colnames(HS)[HS.lut[,2]])
obj <- cophylo(Mul.tree,Lig.tree, linkhs, rotate = TRUE)
## Rotating nodes to optimize matching...
## Done.
plot.cophylo(obj, link.lwd=1, link.lty=1, fsize=0.5, pts=FALSE,
             link.col = "seagreen")

Cophylogentic Signal

We apply PACo to the normalized matrices of patristic distances of Ligophorus spp. and mullets. The significance of the cophylogentic association between hosts and parasites is established based on 1,000 permutations.

pdLig <- LigoMulData$patriticLig/max(LigoMulData$patriticLig)
pdMul <- LigoMulData$patristicMul/max(LigoMulData$patristicMul)
set.seed(2021) # we set a randomization seed to warrant reproducibility.
CS <- prepare_paco_data(pdMul, pdLig, HS)
CS <- add_pcoord(CS, correction = "cailliez")
CS <- suppressWarnings(PACo(CS, nperm = 1000, symmetric = TRUE, 
                             shuffled = TRUE))
CS$gof
## $p
## [1] 0.014
## 
## $ss
## [1] 0.7064206
## 
## $n
## [1] 1000

The analysis returns a m2 of 0.706 and the cophylogenetic association between Ligophorus spp. and grey mullets is significant at the 0.05 level (p = 0.014).

Interaction of Host phylogeny with Ligophorus Shape

We assess the concordance of host phylogeny with the shape of ventral anchors of Ligophorus spp. Ventral anchors are used for attachment to the gills. If significant, one can conclude that host speciation determined anchor shape of Ligophorus spp. Procrustes coordinates of ventral anchors of 244 specimens of Ligophorus (González-Rodríguez et al. (2017) are used to compute the Procrustes distances between species.

LigProc <- LigoMulData$procrustes_coord
LigProc$species <- as.factor(LigProc$species)
splitXspp <- split(LigProc[, 2:17], LigProc$species)
# Compute Procrustes coordinates of species centroids:
spp_centroids <- t(sapply(splitXspp, colMeans))
# Euclidean distances between centroids, i.e. Procrustes distances
PDLigo <- as.matrix(dist(spp_centroids))
# Normalize Procrustes distances
PDLigo <- PDLigo/max(PDLigo)
# Apply PACo
HI <- prepare_paco_data(pdMul, PDLigo, HS)
HI <- add_pcoord(HI, correction = "cailliez")
HI <- suppressWarnings(PACo(HI, nperm = 1000, symmetric = TRUE,
           shuffled = TRUE))
HI$gof
## $p
## [1] 0.067
## 
## $ss
## [1] 0.7915233
## 
## $n
## [1] 1000

PACo yields a m2 of 0.792. The association between mullet phylogeny and shape of ventral anchors of Ligophorus spp. is not significant at the 0.05 level (p = 0.067). The concordance (or rather the lack of thereof) between host phylogeny and parasite shape can be visualized as a tanglegram:

obj <- cophylo(Mul.tree, as.phylo(hclust(dist(spp_centroids))), linkhs,
               rotate = TRUE)
## Rotating nodes to optimize matching...
## Done.
plot.cophylo(obj, link.lwd=1, link.lty=1, fsize=0.5, pts=FALSE,
             link.col = "firebrick")

Interaction of Ligophorus Phylogeny with Mullet Shape

The association between phylogeny of Ligophorus spp. and host morphometry is assessed based on data on fishbase.org. (Ideally information of gill shape should have been used but these data are unavailable.)

# Handle the morphometric data from fishbase.org to get a distance matrix
MulMorph <- LigoMulData$mullets_morph
# Standardize variables & compute Euclidean distances between mullet spp.
MulMorph <- scale(MulMorph)
dMulMorph <- as.matrix(dist(MulMorph))
# Normalize distance matrix
dMulMorph <- dMulMorph/max(dMulMorph)
# Apply PACo
SI <- prepare_paco_data(dMulMorph, pdLig, HS)
SI <- add_pcoord(SI, correction = "cailliez")
SI <- suppressWarnings(PACo(SI, nperm = 1000, symmetric = TRUE,
           shuffled = TRUE))
SI$gof
## $p
## [1] 0.003
## 
## $ss
## [1] 0.6104359
## 
## $n
## [1] 1000

We find a higher concordance between Ligophorus phylogeny and host shape than cophylogenetic signal (m2 = 0.610 vs. m2 = 0.706). This association is highly significant p = 0.003 and again can be inspected with a tanglegram:

obj <- cophylo(as.phylo(hclust(dist(MulMorph))), Lig.tree, linkhs,
               rotate = TRUE)
## Rotating nodes to optimize matching...
## Done.
plot.cophylo(obj, link.lwd=1, link.lty=1, fsize=0.5, pts=FALSE,
             link.col = "steelblue")

Translation to Cophylospace

The plot below shows the position of the system studied in cophylospace. For compatibility with the framework described in the accompanying paper, the axes represent 1 - m2.

plot_ly(x = 1-HI$gof$ss, y = 1-SI$gof$ss, z= 1-CS$gof$ss, type="scatter3d",
        mode="markers") %>%
  layout(
    scene = list(
      xaxis = list(title = "Symbiont interacion", range = c(0,1)),
      yaxis = list(title = "Host interaction", range = c(0,1)),
      zaxis = list(title = "Cophylogenetic signal", range = c(0,1))
    )
  )

Discussion

Under a classical cophylogenetic analysis, the significant congruence between mullet and Ligophorus phylogenies revealed by PACo would indicate that some degree of cospeciation occurred. However, the evidence presented here indicates that morphometric traits of the hosts can explain speciation of Ligophorus better than host phylogeny. The position of our system in cophylospace space (Figure I in Box 2, accompanying paper) suggests some degree of assymmetry in which host traits influenced speciation of Ligophorus spp. In fact, the arrangement of Ligophorus spp. parasitizing Mugil cephalus (L. mediterraneus, L. cephali and L. chabaudi), P. haematocheilus (L. pilengas and L. llewellyni) and C. saliens (L. acuminatus and L. minimus) in respective crown groups seems to indicate several intra-host duplication events. As in other monogeneans (see Blasco-Costa et al. 2012), diversity of microhabitats provided by fish gills could have facilitated intrahost lineage duplications in Ligophorus. This together with high host specificity would explain why phylogenetically related Ligophorus spp. are more likely to interact with the same host species than expected by chance. In any case, cospeciation does not seem to be the major force accounting for speciation of Ligophorus in the Mediterranean and Black Seas.

Funding

Funds were provided by the Swiss National Science Foundation (SNSF grant 31003A_169211 to I. Blasco-Costa) and Ministry of Science and Innovation of Spain (grant PID2019-104908GB-I00 to J.A. Balbuena). A. Hayward is supported by a Biotechnology and Biological Sciences Research Council (BBSRC) David Phillips Fellowship (BB/N020146/1).

References

Balbuena, J.A. et al. (2013) PACo: A Novel Procrustes Application to Cophylogenetic Analysis. PLoS ONE 8, e61048. https://doi.org/10.1371/journal.pone.0061048
Blasco-Costa, I. et al. (2012) Molecular phylogeny of species of Ligophorus (Monogenea: Dactylogyridae) and their affinities within the Dactylogyridae. Parasitology International 61, 619–627. https://doi.org/10.1016/j.parint.2012.06.004
Durand, J.-D. et al. (2012) Systematics of the grey mullets (Teleostei: Mugiliformes: Mugilidae): Molecular phylogenetic evidence challenges two centuries of morphology-based taxonomy. Mol Phylogenet Evol 64, 73–92. https://doi.org/10.1016/j.ympev.2012.03.006
Hutchinson, M.C. et al. (2017) paco: implementing Procrustean Approach to Cophylogeny in R. Methods Ecol Evol 8, 932–940. https://doi.org/10.1111/2041-210X.12736
R Core Team (2021) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
Revell, L.J. (2011) phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol Evol 3, 217–223. https://doi.org/10.1111/j.2041-210X.2011.00169.x
Rodríguez-González, A. et al. (2016) Evolutionary modularity and morphological integration in the haptoral anchor structures of Ligophorus spp. (Monogenea: Dactylogyridae). Parasitol Res 115, 3519–3533. https://doi.org/10.1007/s00436-016-5117-z
Rodríguez-González, A. et al. (2017) Evolutionary morphology in shape and size of haptoral anchors in 14 Ligophorus spp. (Monogenea: Dactylogyridae). PLoS ONE 12, e0178367. https://doi.org/10.1371/journal.pone.0178367
Russo, L. et al. (2017) Quantitative evolutionary patterns in bipartite networks: Vicariance, phylogenetic tracking or diffuse co‐evolution? Methods Ecol Evol 9, 761–772. https://doi.org/10.1111/2041-210X.12914
Sievert C. (2020) Interactive web-based data visualization with R, plotly, and shiny. Chapman and Hall/CRC. ISBN 9781138331457